Key processes in development, including differential cell proliferation, programmed cell death, migration and cell-cell interactions are recapitulated in a pathologic manner during oncogenesis and metastasis, suggesting an aberrant regulation or 're-activation' of such processes. For example, FGF, Wnt, and Shh signaling pathways are all involved in the genesis of certain human neoplasias. Understanding how developmental processes are normally regulated will be invaluable in deciphering tumor biology and ultimately help to identify new ways to intercept cellular targets that drive tumor cell behavior. My lab identified several new transcription factors (homeobox genes and T-box genes) that appear to play roles in the formation and pattern of the primary embryonic axis and the limb axis in vertebrate embryos. Analyzing the role that such regulators play in different developmental contexts may offer new insights into the sorts of basic processes that they govern in the cell. We are analyzing the normal developmental functions of several regulators with the long-term aim of linking steps from initial differences in patterns of gene expression to ultimate differences in tissue morphology and structure in the embryo. In this context, potential roles in oncogenesis and other pathologic conditions are also being considered. 1. Initiation of limb budding: function of the transcription factor T (Brachyury):The initiation of limb bud formation and induction of the apical ectodermal ridge (AER) that directs subsequent limb outgrowth appear to be regulated by a cascade of FGF and Wnt signals. Intermediate target genes in this signal relay are largely unknown. Using retroviral misexpression in chick embryos, we have found that the T-box transcription factor T regulates the formation and maintenance of the AER. T is expressed in the early limb bud (in the mesoderm just underneath the AER) and is regulated by both FGF and Wnt signals. Overexpression causes extension of the AER and additional limb outgrowth, producing extra digits. Mouse embryos mutant for T (-/-) die at E10.5 but do form forelimb buds which are small. These limb buds do not have a normal AER and have very irregular Fgf-8 expression, also consistent with a role for T in normal AER formation. We don't yet know whether T misexpression by itself is sufficient to induce formation of an ectopic limb bud in the flank, as for example, FGF beads or Wnt misexpression can do. Other T-box genes, Tbx5 and Tbx4, regulate both limb identity and limb outgrowth, and also function apparently downstream of Wnt and FGF signals. We are evaluating whether AER formation and outgrowth is regulated by T-Tbx heterodimers and we are generating a conditional knock-down approach to further analyze T function in limb development. In parallel studies, we are developing chromatin immunoprecipitation (ChIP) assays in embryos to allow direct identification of target promoters which T binds to in vivo. We have successfully piloted ChIP assays using Tbx5, for which at least 1 bona fide direct target promoter has been identified (see section 6 below). 2. Determination of limb skeletal pattern: early Hoxd functions in SHH pathway:The limb skeleton arises from a bud by progressive branching of skeletal precursors from proximal to distal (ie. shoulder to hand). The 'pattern' of skeletal elements that form from anterior to posterior (thumb to little finger) is regulated by secreted Sonic hedgehog (SHH) signals from the posterior edge of the limb bud. Hoxd genes are thought to be key targets of SHH signaling that regulate the pattern of skeletal components forming in the limb. Their function in molecular terms and the target genes they regulate are still unknown. Using transgenic mouse models, we found that Hoxd12 regulates Shh expression and the SHH pathway as part of a positive feedback loop in the early limb bud which reinforces and amplifies patterning signals at the right sites. We are evaluating the mechanism by which Hoxd genes can alter the SHH pathway. The Gli3 transcriptional repressor is a major mediator of SHH in the developing limb. We have now found that a direct interaction between Gli3 and Hoxd proteins alters the Sonic hedgehog pathway and skeletal patterning during limb development. Antagonism of Gli3 transcriptional repressor function by SHH signaling derepresses SHH target genes in the limb. Gli3 and Hoxd12 interact genetically and physically, and this interaction converts Gli3 from a repressor into an activator of SHH target genes. Several 5'Hoxd genes interact with Gli3, and we propose that the sum of these interactions determines the levels of expression of Gli3-regulated SHH target genes that regulate the pattern of different digits that form. Recent work has identified a late capability of interdigit signals to change digit identities, even after early digit primordia have already begun to form. These interdigit zones are also late sites of high level Hoxd and Gli3 repressor expression which may thus contribute to this late regulation of digit identity (see sections 4 and 6 below). In total absence of Gli3 (Gli3-/-), digits are small, mainly biphalangeal, despite high Hoxd expression, whereas, in the total absence of SHH (Shh-/-), digit number and size are both severely reduced and NO Gli3 full-length activator is present. Hoxd expression is also very low . We hypothesize that elevation of Hoxd expression will rescue digit development in Shh-/- by tipping the balance of Gli3 repressor:Hoxd levels and are planning Hoxd and Gli3 misexpression experiments in the chick to evaluate the effects on digit patterns formed by varying Hoxd:Gli3 ratios in specific interdigits. We are also devising experiments to alter Gli3-Hoxd ratios at various different times in mice using transgenic approaches. We have crossed a weakly-expressing Hoxd12 transgenic line with Shh mutant mice to generate Shh-/- embryos with modestly increased Hoxd12 expression. Even in the presence of very high Gli3 repressor (Gli3+/+), the size and number of phalangeal segments are increased in the single digit that develops without Shh when Hoxd12 levels are raised. This rescue of more 'normal' digit morphologies is further enhanced by simultaneously reducing levels ofGli3 repressor (Gli3 +/-) and raising Hoxd12 levels transgenically in Shh-/- embryos. We are currently analyzing limb-bud stage embryos to define important signaling pathways that are altered when Hoxd12 level is raised in Shh-/- embryos (particularly Bmp, Wnt). These pathways are likely to lie downstream of Gli3-Hoxd interaction. 3. Chondrogenic differentiation of limb skeletal elements: late Hoxd functions: Hoxd genes continue to be expressed quite late at the periphery of chondrogenic condensations that will form the distal skeletal elements (digits), as well as certain long bone progenitors at an earlier stage, but their expression generally shuts off as cartilage differentiation proceeds within the condensations. There is evidence that the early Hoxd expression drives expansion of undifferentiated mesoderm (allowing more digits to eventually form). Is their shut-off important to allow the transition from proliferation to differentiation to proceed? We developed an inducible transgenic model to selectively analyze later developmental functions of 5' Hoxd genes in association with condensing and differentiating cartilage. This conditional expression approach (using Cre recombinase) will also allow us to identify molecular targets using DNA microarray analysis of transgenic embryonic limb buds. Our preliminary results with late Hoxd12 or Hoxd13 misexpression indicate that Hoxd gene expression must shut off to allow early steps in cartilage differentiation for cartilage precursors of only certain bone elements.